Merge branch 'groupby' into 'master'

Added HBaseMimeCount{Job,Test}, which counts the number of rows with each mimetype, as an example of groupby

See merge request webgroup/sandcrawler!4

4 files changed, 178 insertions, 0 deletions

diff --git a/scalding/src/main/scala/sandcrawler/HBaseMimeCountJob.scala b/scalding/src/main/scala/sandcrawler/HBaseMimeCountJob.scala
new file mode 100644
index 0000000..819a652
--- /dev/null
+++ b/scalding/src/main/scala/sandcrawler/HBaseMimeCountJob.scala
@@ -0,0 +1,28 @@
+package sandcrawler
+
+import cascading.property.AppProps
+import cascading.tuple.Fields
+import com.twitter.scalding._
+import java.util.Properties
+import parallelai.spyglass.base.JobBase
+import parallelai.spyglass.hbase.{HBaseSource, HBasePipeConversions}
+import parallelai.spyglass.hbase.HBaseConstants.SourceMode
+
+class HBaseMimeCountJob(args: Args) extends JobBase(args) with HBasePipeConversions {
+  val output = args("output")
+
+  HBaseMimeCountJob.getHBaseSource
+    .read
+    .fromBytesWritable(List('mime))
+    .debug
+    .groupBy('mime){group => group.size('count)}
+    .write(Tsv(output))
+}
+
+object HBaseMimeCountJob {
+  def getHBaseSource = HBaseBuilder.build(
+    "wbgrp-journal-extract-0-qa",     // HBase Table Name
+    "mtrcs-zk1.us.archive.org:2181",  // HBase Zookeeper server (to get runtime config info; can be array?)
+    List("file:mime"),
+    SourceMode.SCAN_ALL)
+}
diff --git a/scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala b/scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala
new file mode 100644
index 0000000..0675efc
--- /dev/null
+++ b/scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala
@@ -0,0 +1,28 @@
+package sandcrawler
+
+import cascading.property.AppProps
+import cascading.tuple.Fields
+import com.twitter.scalding._
+import java.util.Properties
+import parallelai.spyglass.base.JobBase
+import parallelai.spyglass.hbase.{HBaseSource, HBasePipeConversions}
+import parallelai.spyglass.hbase.HBaseConstants.SourceMode
+
+class HBaseStatusCountJob(args: Args) extends JobBase(args) with HBasePipeConversions {
+  val output = args("output")
+
+  HBaseStatusCountJob.getHBaseSource
+    .read
+    .fromBytesWritable(List('status))
+//    .debug
+    .groupBy('status){group => group.size('count)}
+    .write(Tsv(output))
+}
+
+object HBaseStatusCountJob {
+  def getHBaseSource = HBaseBuilder.build(
+    "wbgrp-journal-extract-0-qa",     // HBase Table Name
+    "mtrcs-zk1.us.archive.org:2181",  // HBase Zookeeper server (to get runtime config info; can be array?)
+    List("grobid0:status"),
+    SourceMode.SCAN_ALL)
+}
diff --git a/scalding/src/test/scala/sandcrawler/HBaseMimeCountTest.scala b/scalding/src/test/scala/sandcrawler/HBaseMimeCountTest.scala
new file mode 100644
index 0000000..eb6f4ff
--- /dev/null
+++ b/scalding/src/test/scala/sandcrawler/HBaseMimeCountTest.scala
@@ -0,0 +1,61 @@
+package sandcrawler
+
+import cascading.tuple.{Tuple, Fields}
+import com.twitter.scalding.{JobTest, Tsv, TupleConversions}
+import org.apache.hadoop.hbase.io.ImmutableBytesWritable
+import org.apache.hadoop.hbase.util.Bytes
+import org.junit.runner.RunWith
+import org.scalatest.FunSpec
+import org.scalatest.junit.JUnitRunner
+import org.slf4j.LoggerFactory
+import parallelai.spyglass.hbase.HBaseSource
+import parallelai.spyglass.hbase.HBaseConstants.SourceMode
+import scala._
+
+@RunWith(classOf[JUnitRunner])
+class HBaseMimeCountTest extends FunSpec with TupleConversions {
+
+  val output = "/tmp/testOutput"
+
+  val log = LoggerFactory.getLogger(this.getClass.getName)
+
+  val mimeType1 = "text/html"
+  val mimeType2 = "application/pdf"
+
+  val sampleData = List(
+    List("sha1:K2DKSSVTXWPRMFDTWSTCQW3RVWRIOV3Q", mimeType1),
+    List("sha1:C3YNNEGH5WAG5ZAAXWAEBNXJWT6CZ3WU", mimeType1),
+    List("sha1:SDKUVHC3YNNEGH5WAG5ZAAXWAEBNX4WT", mimeType2),
+    List("sha1:35985C3YNNEGH5WAG5ZAAXWAEBNXJW56", mimeType2),
+    List("sha1:885C3YNNEGH5WAG5ZAAXWA8BNXJWT6CZ", mimeType2),
+    List("sha1:00904C3YNNEGH5WAG5ZA9XWAEBNXJWT6", mimeType2),
+    List("sha1:249C3YNNEGH5WAG5ZAAXWAEBNXJWT6CZ", mimeType1),
+    List("sha1:095893C3YNNEGH5WAG5ZAAXWAEBNXJWT", mimeType2)
+  )
+
+  val mimeType1Count = sampleData.count(lst => lst(1) == mimeType1)
+  val mimeType2Count = sampleData.count(lst => lst(1) == mimeType2)
+
+  JobTest("sandcrawler.HBaseMimeCountJob")
+    .arg("test", "")
+    .arg("app.conf.path", "app.conf")
+    .arg("output", output)
+    .arg("debug", "true")
+    .source[Tuple](HBaseMimeCountJob.getHBaseSource,
+      sampleData.map(l => new Tuple(l.map(s => {new ImmutableBytesWritable(Bytes.toBytes(s))}):_*)))
+      .sink[Tuple](Tsv(output)) {
+        outputBuffer =>
+        it("should return a 2-element list.") {
+          assert(outputBuffer.size === 2)
+        }
+
+        // Convert List[Tuple] to Map[String, Integer].
+        val counts = outputBuffer.map(t => (t.getString(0), t.getInteger(1))).toMap
+        it("should have the appropriate number of each mime type") {
+          assert(counts(mimeType1) == mimeType1Count)
+          assert(counts(mimeType2) == mimeType2Count)
+        }
+    }
+    .run
+    .finish
+}
diff --git a/scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala b/scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala
new file mode 100644
index 0000000..8b5c3d6
--- /dev/null
+++ b/scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala
@@ -0,0 +1,61 @@
+package sandcrawler
+
+import cascading.tuple.{Tuple, Fields}
+import com.twitter.scalding.{JobTest, Tsv, TupleConversions}
+import org.apache.hadoop.hbase.io.ImmutableBytesWritable
+import org.apache.hadoop.hbase.util.Bytes
+import org.junit.runner.RunWith
+import org.scalatest.FunSpec
+import org.scalatest.junit.JUnitRunner
+import org.slf4j.LoggerFactory
+import parallelai.spyglass.hbase.HBaseSource
+import parallelai.spyglass.hbase.HBaseConstants.SourceMode
+import scala._
+
+@RunWith(classOf[JUnitRunner])
+class HBaseStatusCountTest extends FunSpec with TupleConversions {
+
+  val output = "/tmp/testOutput"
+
+  val log = LoggerFactory.getLogger(this.getClass.getName)
+
+  val statusType1 = "200"
+  val statusType2 = "404"
+
+  val sampleData = List(
+    List("sha1:K2DKSSVTXWPRMFDTWSTCQW3RVWRIOV3Q", statusType1),
+    List("sha1:C3YNNEGH5WAG5ZAAXWAEBNXJWT6CZ3WU", statusType1),
+    List("sha1:SDKUVHC3YNNEGH5WAG5ZAAXWAEBNX4WT", statusType2),
+    List("sha1:35985C3YNNEGH5WAG5ZAAXWAEBNXJW56", statusType2),
+    List("sha1:885C3YNNEGH5WAG5ZAAXWA8BNXJWT6CZ", statusType2),
+    List("sha1:00904C3YNNEGH5WAG5ZA9XWAEBNXJWT6", statusType2),
+    List("sha1:249C3YNNEGH5WAG5ZAAXWAEBNXJWT6CZ", statusType1),
+    List("sha1:095893C3YNNEGH5WAG5ZAAXWAEBNXJWT", statusType2)
+  )
+
+  val statusType1Count = sampleData.count(lst => lst(1) == statusType1)
+  val statusType2Count = sampleData.count(lst => lst(1) == statusType2)
+
+  JobTest("sandcrawler.HBaseStatusCountJob")
+    .arg("test", "")
+    .arg("app.conf.path", "app.conf")
+    .arg("output", output)
+    .arg("debug", "true")
+    .source[Tuple](HBaseStatusCountJob.getHBaseSource,
+      sampleData.map(l => new Tuple(l.map(s => {new ImmutableBytesWritable(Bytes.toBytes(s))}):_*)))
+      .sink[Tuple](Tsv(output)) {
+        outputBuffer =>
+        it("should return a 2-element list.") {
+          assert(outputBuffer.size === 2)
+        }
+
+        // Convert List[Tuple] to Map[String, Integer].
+        val counts = outputBuffer.map(t => (t.getString(0), t.getInteger(1))).toMap
+        it("should have the appropriate number of each status type") {
+          assert(counts(statusType1) == statusType1Count)
+          assert(counts(statusType2) == statusType2Count)
+        }
+    }
+    .run
+    .finish
+}